Hippocampal Dynamics Underlying Memory Formation and Consolidation
Memory consolidation is thought to involve a hippocampo-cortical dialogue during sleep to stabilize labile memory traces for long-term storage. Over the recent years, we and others have provided both direct causal evidence for this theory, and key insights into the underlying network dynamics.
Hippocampal ‘place’ cells code for the location of the animal in the environment: they selectively discharge in specific locations in the environment, referred to as ‘firing fields’. As a rat crosses successive firing fields, the corresponding place cells are maximally activated one after the other. As a result, their activity represents the ongoing trajectory. Note that the trajectory is represented in real time, i.e. at the behavioral timescale. However, because successive firing fields overlap with each other, at a faster timescale the respective spike trains are actually intermingled. And strikingly, careful examination of this collective activity reveals that within each individual cycle of the ongoing theta rhythm (~8 Hz), place cells also tend to activate in the same order, again representing the ongoing trajectory - but at a highly accelerated (5-10x) speed.
There are at least three reasons to suspect that hippocampal sequences are related to memory processes. First, these sequences are repeated at theta frequency, which dramatically increases the number of times they occur in the hippocampal network. And synaptic plasticity is facilitated by repeated co-firing. Second, they allow hippocampal neurons to fire in close temporal proximity (dozens of milliseconds apart, compared to hundreds of milliseconds at the behavioral timescale). And classical plasticity mechanisms require brief delays between successively activated neurons. Third, theta sequences represent the past, present and future locations of the rat. Hence, they include two critical components (space and time) of episodic memory.
Perhaps even more remarkably, the same sequences are later replayed during slow wave sleep — as if the rats were ‘dreaming’, or virtually reexploring the same environment in thought. This takes place during brief high frequency oscillations (‘ripples’) and has been hypothesized to support memory consolidation, possibly via a dialogue with the neocortex.
Our work addresses the following questions: How are theta sequences initially generated during exploration? Are theta sequences required to form initial memory traces? Is subsequent replay during sleep necessary for memory consolidation? Does hippocampus replay trigger functional reorganization in cortical networks? Does cortical activity subsequently reflect such reorganization during memory recall? Is this hippocampo-cortical dialogue the underlying network mechanism of memory consolidation?
Centre interdisciplinaire de recherche en biologie – Collège de France, Paris
20 février 2018 – 11 h
Inmed – Salle de Conférence